Current location - Trademark Inquiry Complete Network - Trademark registration - E-commerce platform product design: splitting the front and backend categories of products
E-commerce platform product design: splitting the front and backend categories of products

First, let’s talk about the splitting of front-end and back-end categories. If we delve into history, we will find that the separation of front-end and back-end categories was first done on the old Taobao platform. The earliest Tmall used a compromise solution. By embedding a large number of editable content modules in the List/search page, the waiter can edit at any time to form a shopping guide path. However, this does not affect the idea of ??analysis.

Recall the conflict between the waiter and the seller: the waiter keeps adjusting the category attributes for the sake of shopping guide; the seller does not want to adjust for the stability of the product. The difference is actually the conflict between the needs of buyers and sellers. The waiter is also in a dilemma. On the one hand, the shopping guide is to give buyers a better experience, but on the other hand, he hopes that the sellers will not have to deal with it all the time.

The essence here is that one product cannot satisfy two completely different user groups well. What should we do? It’s a very simple logic. Split this product into two products, one to satisfy buyers and the other to satisfy sellers. How to do it specifically? You can understand it by going to a big supermarket.

The way the products are displayed on the shelves in large supermarkets changes, and is often adjusted with the seasons, sales, and specific events. There is also a warehouse in a large supermarket. The placement of goods in the warehouse is relatively fixed. Food is in the food area, and washing and care is in the washing and care area. Therefore, products are actually placed in two places - front-end sales and back-end storage, corresponding to two classification methods. You should understand that in many cases, e-commerce lags behind traditional commerce. After hundreds of years of development, there must be many things that are "better" than us, and we need to learn more. In this way, we split the front-end and back-end categories into front-end categories and back-end categories.

To put it simply: sellers publish products through backend categories, and buyers select products through frontend categories.

The original category has become a backend category. It is still a tree, and its attributes are still attributes. Products are all listed in the backend categories and are listed according to category attributes.

To put it simply, the front-end category method is to first build a category tree according to the requirements of the front-end shopping guide, and then establish an association between the leaf categories of this tree and the back-end categories. Any front-end leaf category can correspond to any one or more back-end categories, and it does not have to be a back-end leaf category. Attribute filtering can be added, and it can be directly associated with SPU, and various product filtering conditions can be added. For example, buyer’s credit, title keywords, whether there is consumer protection, etc.

Use examples to illustrate.

For example, for women's clothing, it's summer now. I hope to adjust the miniskirt to be easier to see. I can build a front-end category of "women's clothing-miniskirt", and then directly add the four categories to the back-end category. Level 5 miniskirts are hung directly under the miniskirts in the front desk category.

For example, if Jack and Jones sells very well, you can create a front-end leaf category of Jack and Jones. You can go to each back-end category of clothing and add the brand Jack and Jones. There may be 10 categories. The current Jack Jones brand is hung under the Jack Jones at the front desk, which means that all Jack Jones men's clothing, women's clothing, children's clothing, accessories, etc. are all together.

For example, if the iPhone 5S sells very well, the waiter can create a front-end category for iPhone 5S, and then go to the back category and add the only SPU with the brand iPhone and model 5S to this front-end category. Down.

Let’s take another look at the picture below.

The light gray area in the upper right corner is the backend category. The brand, gender, suitable crowd, suitable occasion, etc. in slightly darker colors below are attributes. Below the attributes are attribute values. Nike (NIKE), casual, Long sleeves etc. The dark gray area on the upper left is the front desk category. I can build a personalized category such as "Vibrant Youth-Street Sports" on the front desk, and then hang the products under the back-end category "Women's Clothing-T-Shirts" with "Brand=NIKE&Gender=Male&Suitable Group=Teenagers" on the front desk. under category.

Let me emphasize again that the backend categories can only be hung under the leaf categories of the foreground categories. In fact, for the front-end category, it generally does not exceed the third level. Because of the length of the shopping guide path, it is meaningless to exceed the third level.

In order to better guide shopping, the concept of public attributes has also been used in front-end categories. If all sub-categories under a front-end parent category have unified attributes, then the attributes of the sub-categories can be extracted as Public *** attributes. It is a pity that by the beginning of 2012, this action still needs people to do it. The waiter needs to press a button to extract the attributes of the company. I hope that subsequent product managers can make improvements.

The benefits of splitting categories into front-end and back-end categories are obvious. A relatively fixed back-end category can be used to derive a front-end category that changes at any time. The back-end category satisfies sellers, and the front-end category satisfies buyers. Everyone is happy. Later, Taobao, Tmall, Electronic City, etc. all derived different front-end categories based on the same set of back-end categories to meet different needs.

But there is a small "dark cloud" in the clear sky - the buyer is at the front desk, looking for products through the front desk categories, and the front desk categories are built by the waiter's "human flesh", so it is very Maybe, some products under the backend categories are not listed under any frontend leaf categories, which means that some products cannot be found through the category shopping guide (of course, it has no impact on the search).

Fortunately, our product manager quickly drove away this "dark cloud". We provided reminders to the operations assistants and told them what the "backend category coverage rate" was. But this is not enough. It just reminds them that good tools have not been provided to them. This is also an optimization worth looking forward to.

Returning to the picture above, there are still a lot of old concepts below, but there are new things for Taobao waiting for us to digest.

10. Old new concepts: SPU/SKU and others?

In e-commerce, SPU (sometimes also called product), SKU (Stock Keeping Unit, stock keeping unit) , product, etc. If you can’t distinguish these words, you will be embarrassed to say hello to others, so let’s take a look at SPU first.

The Tmall platform has made many subdivisions of the attributes of the backend categories. There is an attribute called "key attribute", such as clothing, brand and item number. The combination of these two attribute values ??can form A unique product - SPU. SPU is equivalent to a standardized product. By determining the key attributes, an SPU is determined. Apple's iPhone 5S is an SPU.

At that time, our idea was very simple. The Taobao platform may have many non-standardized products, but as a B2C platform, it is best for all products to be standardized, which makes it easier for us to manage. Therefore, we feel that all products should correspond to an SPU. Although there are various types of clothing, from the manufacturer, a style is a style and a model is a model. They must be standardized and must have SPU. For a certain type of clothing, the manufacturer has never produced shit yellow. There should be no seller selling this color. All products have corresponding SPU, which means that the basic information is available. After the SPU is defined, for example, if he knows a pair of shoes, the brand is Nike and the item number is 12345, this is the only SPU. Other attribute values ??are hung on the SPU and will not change due to changes in sellers. They are fixed and standardized.

There are products hanging under the SPU, such as a pair of Nike 12345. Five merchants are selling them, that is, five products. The difference between the product and the SPU is that it is related to the seller. Products sandwiched between SPU and SKU are a concept unique to B2C platforms. If it is an independent B2C, there is no situation where an SPU is sold by multiple sellers, and it can be directly divided into SKUs.

In order to explain SKU, we have to look at how the concept of "sales attributes" was proposed. In a seemingly complex system, one concept after another is proposed to solve actual business problems.

In the past, when Taobao buyers bought a piece of clothing, they had to leave a note to the merchant saying that they wanted "black" and "XL size" clothing. When sellers send a product, they can only send clothes such as Li-Ning brand and item number 56789. One item has several colors and sizes, but there is only one inventory, so all the sizes are combined. The inventory is all combined. In fact, there are only 10 pieces of black large size clothes, but because the total inventory is 100 pieces, if 12 users buy the black large size at this time, the latter two will not be in stock. The solution is sold.

This is actually the concept of inventory management and its application in real-life scenarios. When your transaction volume is not large, these are easy to solve. You can say: "I'm sorry, it's gone. Refund, can you exchange it?" The bigger size will be more spacious and cooler. "But if you want to shop independently on the B2C platform, this cannot be solved. Therefore, we need to continue to classify and refine the products, so we put forward the concept of sales attributes. Originally, this was a very normal classification method offline, but in 2007 Taobao product managers had little offline experience, so they thought of a concept called "minimum purchase unit". Later we found that it was different from offline The SKU - Standard Stock Keeping Unit - is actually the same thing.

Our choice at that time was to make the SKU as simple as possible. We believed that whether the iPhone 5S is white or black, it is only an attribute difference with the SPU, not an essential difference, so there is no Think of them as two SPUs. Color, size, etc. have become sales attributes now.

(Note: At the beginning of 2012, Tmall proposed SKU-based SPU. Simply put, SKUs that have been segmented by sales attributes are managed as SPUs, so that SPUs are consistent with offline inventory management. . Students who are interested in learning more may wish to join Tmall. )

Finally, we would also like to mention the concept of "product attributes", which include some attributes, such as condition, warranty, additional services, etc. , it has nothing to do with SKU or SPU. For example, for people who buy mobile phones, as long as they buy mobile phones in China, there are differences in warranty. Some mobile phones are covered by a nationwide joint warranty, while other mobile phones only provide store warranty. This attribute is a product attribute and has nothing to do with SPU or SKU. It has something to do with the seller. Although only a few attributes are commodity attributes, we propose this concept so that the entire commodity system is relatively complete.

Okay, that’s enough about the special new things on the Tmall platform. Do you still remember the miserable Taobao platform? Another year or two have passed, and the problems on that platform have become more and more prominent.

Just like that, the time has arrived in 2009.

Tmall has made a lot of innovations in terms of products and business. However, the transaction volume has not increased. At that time, it was 2 million per day. When the first general manager resigned, it was less than 4 million. Basically, There is not much growth, why? Tmall is an independent website, and there are not enough products to attract buyers. If buyers do not come, there will be no transaction. If there is no transaction, the company will have no patience. If there is no patience, the general manager will leave. Taobao Mall was originally a business unit. This business unit was disbanded in less than a year. Those who recruited investment were reorganized into investment promotion, and those who operated were reorganized into operations. They all returned to Taobao.

Almost three months later, the company couldn't sit still anymore. It felt that we couldn't give up on B2C. How could we give up? At that time, we discussed why we couldn't do it. The important conclusion was that Taobao Mall was By developing independently, there is no way to fully enjoy Taobao's existing advantages - users and traffic. Based on this guiding ideology at that time, the senior management decided to close down B2C, which here refers to the underlying systems such as commodities and transactions. A Jian was the product manager who was most involved in the entire process. He set up Tmall in 2007 and closed it down in 2009.

Throughout 2009, we merged Tmall and Taobao. But there were two options at the time. The initial idea was to move the mall data back to Taobao. The second one was Ajian’s suggestion, saying that it would be a pity to dismantle it like this (this is a product he made himself, and he worked hard to build the entire mall, so it would be a pity to abolish it). Can Taobao be upgraded to Taobao? The mall is like this and then closed. Three months is enough for the first legalization, and half a year may be needed for the later legalization. There was a lot of controversy at the time. Some supported the first option, while others supported the second option. In the end, one person made the key decision, and that was Lao Lu - Temujin (the second boss of Taobao, who succeeded Jack Ma to become Alibaba in 2013). CEO of the group).

The following is gossip, dictated by Ajian:?

Not long after Lao Lu came to Taobao, I had a meeting with him, and he made some suggestions there. I felt that Lao Lu didn’t understand anything. Like, I wonder if this CEO is unreliable. But when I heard Lao Lu say, "How can you give up such a good experience in the mall?", my view completely changed. At a critical moment, as a CEO, Lao Lu made a crucial choice.

The most important thing for a CEO is to make key decisions at critical moments. I immediately felt that he was a very qualified CEO. BC integration is to move the mall system to Taobao, which is Taobao's largest project that year - Colorful Stone (this project will be mentioned many times later).

In the entire Taobao, half of the development revolves around a product manager. It is estimated that there will be no experience in the future. This is a rare opportunity for a product manager.

After this system was established, the entire Taobao system has basically remained unchanged. Even in June 2011, Taobao and Tmall were separated in terms of organizational structure, and this was the case in mid-2012. There are some other things worth saying, let’s list them.

For many years, Tmall required sellers to publish products under an existing SPU. Therefore, when the SPU did not exist or the seller could not find it, it would require The seller interrupts the product publishing process and publishes SPU first. This problem troubled the seller for many years and was only solved in 2011.

In fact, the reason behind this is not whether it can be solved technically, but the issue of the management scale of the B2C platform, whether it is loose or tight, and where this level is, this is something we have been thinking about. Now it seems that the previous assumptions were too idealistic and the system was too rigid. Sellers should not have so many restrictions on releasing products. As a platform, although the management of goods should be as standardized as possible, everything should not be imposed on sellers.

In the early days, our idea was very simple. For an Apple iPhone 5S to undergo product quality inspection, it must have an SPU. We trust every merchant to issue a standard SPU. Once this product is distributed, other merchants do not need to issue it. So it means that the first merchant to ship iPhone 5S must first release an SPU, and then only the product will be released.

Allow him to issue SPU. In fact, for non-standardized categories such as clothing, it basically means that each product must be issued with an SPU. Therefore, when a seller sends a product, he must first send an SPU, which is very troublesome and painful for the seller. Therefore, we will combine these two processes later. When a buyer sends a product, we will judge whether there is this SPU: if there is this SPU, then the product will be sent directly; if there is no SPU, an SPU will be automatically attached. Go up.

There is another episode here. In the era when SPU was released first and then the products, for a period of time, we found that some merchants were really unreliable and the SPU they sent was full of errors, so we still sold them in some categories. The concept of "trusting merchants" is proposed, that is, some excellent merchants are identified and marked, and only these merchants can issue SPU. This creates a more troublesome problem - untrusted merchants may not be able to ship products because the corresponding trusted merchants have not yet had time to ship SPU.

This is a question of control scale. Finally, let’s talk about in-depth thinking on product classification. We have always said that the core of product management is classification, so broadly speaking, merchants, vertical markets, etc. are also differentiating dimensions.

However, the vertical market is not a simple classification. It is based on the characteristics of the goods and distinguishes information that cannot be provided by the classification, that is, services. In terms of service standards, if you just classify mobile phones, if this mobile phone vertical market is no different from other categories, there is no need to call it a vertical market, just call it a mobile phone category. Only when the mobile phone product itself has other service features, we need to build a vertical market, such as extended warranty.

Similar to "consumer protection", the action of marking products is a similar concept. For example, consumer protection is a cross-category purpose, and the “compensation for one fake product” may only apply to cosmetics.

Product managers who are most familiar with the entire Taobao product system believe that the biggest flaw of the product system is that it is very rigid. A system that is too rigorous is not necessarily a good thing. It will make it difficult for new things to emerge and evolve. Taobao is actually a platform that needs a hundred flowers to bloom in an orderly manner.

This is the fifth generation of product management that Taobao has introduced internally since 2012. Its key words are: search, tagging, and scenario-based. However, this is the latest thing that happened, and no one can see clearly. Where the future leads is determined by product managers themselves, we will wait and see.

Everyone thinks there is a possibility for the development direction of the entire product management, which is search. Sellers post products in an orderly manner. Powerful search is the classification of products.